Understanding Predictive and Holistic Insights in WebAdMIT

Overview

Liaison’s predictive and holistic insights for WebAdMIT, powered by Othot’s data science models, introduce a more comprehensive way to evaluate applicants by combining predictive analytics with holistic review. Together, these models provide two distinct but complementary perspectives: one grounded in historical enrollment patterns and one focused on the full context of an applicant’s profile.

This article provides foundational information to help you understand predictive and holistic insights within WebAdMIT. These features are available only to participating institutions and CASs. For guidance on how to view, configure, and use these scores within WebAdMIT – including adjusting weights and incorporating scores into workflows – refer to the corresponding help articles on accessing and using scores.

Understanding Predictive and Holistic Models

Predictive and holistic models are designed to answer different – but equally important – questions about your applicant pool.

The holistic model evaluates each applicant based on the full scope of their application, including academics, experiences, attributes, and context. It is designed to reflect how well an applicant aligns with your institution’s mission, values, and priorities.
The predictive model uses historical data and patterns to estimate the likelihood that an applicant will matriculate if admitted. By analyzing trends across previous cycles, it provides insight into expected enrollment behavior and yield.

Individually, each model offers value. Together, they provide a more complete view – combining qualitative alignment with data-informed projections.

Utilizing Predictive & Holistic Insights Together

Together, predictive and holistic models provide complementary perspectives on each applicant, helping institutions prioritize outreach, tailor decision strategies, and better understand the composition and needs of their overall applicant pool.

High Scores in Both Models

Applicants who score highly in both models demonstrate strong alignment with institutional priorities and a high likelihood of matriculation. They often exhibit the preparation, experiences, and attributes programs seek, and they are more likely to enroll if admitted.
Because these applicants support both mission alignment and enrollment outcomes, they are typically strong candidates for admission and may be prioritized accordingly.

Low Scores in Both Models

When both the predictive and holistic scores are low, these applicants may show less alignment with institutional priorities and also have a lower likelihood of matriculation.
In a competitive pool, this group is often a lower priority for both admission and outreach, allowing programs to focus time and resources on applicants with stronger overall alignment.

Low Predictive, High Holistic

Applicants with high holistic scores continue to reflect many of the qualities an institution values, including strong preparation and alignment with mission. However, a lower predictive score suggests they may be less likely to matriculate – often because they are highly competitive candidates with multiple attractive options.
Many programs still choose to admit these applicants, recognizing their potential value to the class. At the same time, institutions must weigh the likelihood of lower yield against the importance of attracting high-potential candidates, making this a deliberate and strategic trade-off.

High Predictive, Low Holistic

When the predictive score is high but the holistic score is lower, the data suggests a strong likelihood of matriculation, but less alignment with the program’s qualitative criteria or mission.
These applicants may have fewer competing offers and therefore enroll at higher rates when admitted, which can be appealing from a yield perspective. However, institutions must balance these benefits against potential impacts on class composition, student success, and long-term goals.

Availability & Getting Started

Predictive and holistic insights may be made available through your CAS association partner. Institutions that choose to participate work with Liaison to implement and configure the models within WebAdMIT. As part of onboarding, institutions receive guidance on the setup process, including data access, configuration requirements, and institution-specific customization options.

For more information about availability or implementation options, contact a member of your account team.

Institution Implementation Process

Once an institution elects to implement predictive and/or holistic insights, Liaison works with the institution to configure data access, implement the model, and integrate scores into WebAdMIT. The following outlines the high-level implementation process.

Step 1: Provide Data Access

To support model development, your institution provides Liaison with appropriate access to WebAdMIT through a dedicated user account with the necessary permissions. This access allows Liaison to securely configure data exports and integrate predictive and holistic insights back into your system.

Access is granted for the application cycles and programs relevant to your modeling goals. For holistic models, this typically includes the current cycle. Predictive models generally require an additional three years of historical data to identify patterns and train accurate enrollment models.

Step 2: Configure Data Exports

Once access is established, Liaison configures data exports that include the standard set of information used to build enrollment models. These exports draw from a broad range of applicant data, including:

Applicant demographics
Academic background (e.g., institutions attended, GPAs)
Test scores (e.g., MCAT)
Experiences and activities (e.g., employment, service, extracurriculars)
Application details (e.g., designations, personal statements, fee waivers)
Geographic information
Program-specific indicators

This feature set reflects data elements that have consistently proven valuable across enrollment modeling efforts. Specific fields and configurations may vary based on your CAS type.

Step 3: Review and Validate Data

Liaison collects data from the configured exports for all relevant cycles and securely transfers it for processing and analysis. As part of this process, the data is reviewed for completeness and quality. Any gaps, inconsistencies, or questions related to data definitions are identified and reviewed with your institution to ensure the data is accurately interpreted before model development begins.

Step 4: Receive Ongoing Model Updates

Once the models are finalized, Liaison integrates predictive and holistic insights directly into WebAdMIT. This integration runs on a regular (daily) schedule, ensuring that scores and related data remain current as new applications are submitted and applicant information is updated throughout the cycle.